• The KIPS Transactions:PartD
• /
• v.13D no.7 s.110
• /
• pp.857-866
• /
• 2006

Performance of an image retrieval system is usually very low because of the semantic gap between the low level feature and the high level concept in a query image. Semantically relevant images may exhibit very different visual characteristics, and may be scattered in several clusters. In this paper, we propose a content based image rertrieval approach which combines region based image retrieval and a new relevance feedback method using adaptive clustering together. Our main goal is finding semantically related clusters to narrow down the semantic gap. Our method consists of region based clustering processes and cluster-merging process. All segmented regions of relevant images are organized into semantically related hierarchical clusters, and clusters are merged by finding the number of the latent clusters. This method, in the cluster-merging process, applies r: using v principal components instead of classical Hotelling's $T_v^2$ [1] to find the unknown number of clusters and resolve the singularity problem in high dimensions and demonstrate that there is little difference between the performance of $T^2$ and that of $T_v^2$. Experiments have demonstrated that the proposed approach is effective in improving the performance of an image retrieval system.

• The Journal of the Korea Contents Association
• /
• v.4 no.4
• /
• pp.163-169
• /
• 2004

In this paper. we propose a shaded recognition method using symmetric feature. When the existing PCA is applied to shaded face images, the recognition rate is decreased. To improve the recognition rate, we use facial symmetry. If the difference of light and shade is greater than a threshold value, we make a mirror image by replacing the dark side with the bright side symmetrically Then the mirror image is compared with a query image. We compare the performance of the proposed algorithm with the existing algorithms such as PCA, PCA without three eigenfaces and histogram equalization methods. The recognition rate of our method shows $98.889\%$ with the excellent result.

• Journal of Korea Multimedia Society
• /
• v.17 no.1
• /
• pp.26-33
• /
• 2014

Video retrieval and indexing are performed by comparing feature similarities between key frames in shot after detecting a scene change and extracting key frames from the shot. Typical image features such as color, shape, and texture are used in content-based video and image retrieval. Many approaches for integrating these features have been studied. However, the issue of these approaches is how to appropriately assign weighting of key frame features at query time. Therefore, we propose a new video retrieval method using adaptively weighted image features. We performed computer simulations in test databases which consist of various kinds of key frames. The experimental results show that the proposed method has better performance than previous works in respect to several performance evaluations such as precision vs. recall, retrieval efficiency, and ranking measure.

• Journal of KIISE:Software and Applications
• /
• v.26 no.10
• /
• pp.1204-1218
• /
• 1999

문서영상 검색을 위한 디지털도서관의 대부분은 논문제목과/또는 논문요약으로부터 만들어진 색인에 근거한 제한적인 검색기능을 제공하고 있다. 본 논문에서는 영문 문서영상전체에 대한 검색을 위한 단어 영상 형태 특징기반의 단어검색시스템을 제안한다. 본 논문에서는 검색의 효율성과 정확도를 높이기 위해 1) 기존의 단어검색시스템에서 사용된 특징들을 조합하여 사용하며, 2) 특징의 개수 및 위치뿐만 아니라 특징들의 순서를 포함하여 매칭하는 방법을 사용하며, 3) 특징비교에 의해 검색결과를 얻은 후에 여과목적으로 문자인식을 부분적으로 적용하는 2단계의 검색방법을 사용한다. 제안된 시스템의 동작은 다음과 같다. 문서 영상이 주어지면, 문서 영상 구조가 분석되고 단어 영역들의 조합으로 분할된다. 단어 영상의 특징들이 추출되어 저장된다. 사용자의 텍스트 질의가 주어지면 이에 대응되는 단어 영상이 만들어지며 이로부터 영상특징이 추출된다. 이 참조 특징과 저장된 특징들과 비교하여 유사한 단어를 검색하게 된다. 제안된 시스템은 IBM-PC를 이용한 웹 환경에서 구축되었으며, 영문 문서영상을 이용하여 실험이 수행되었다. 실험결과는 본 논문에서 제안하는 방법들의 유효성을 보여주고 있다. Abstract Most existing digital libraries for document image retrieval provide a limited retrieval service due to their indexing from document titles and/or the content of document abstracts. This paper proposes a word spotting system for full English document image retrieval based on word image shape features. In order to improve not only the efficiency but also the precision of a retrieval system, we develop the system by 1) using a combination of the holistic features which have been used in the existing word spotting systems, 2) performing image matching by comparing the order of features in a word in addition to the number of features and their positions, and 3) adopting 2 stage retrieval strategies by obtaining retrieval results by image feature matching and applying OCR(Optical Charater Recognition) partly to the results for filtering purpose. The proposed system operates as follows: given a document image, its structure is analyzed and is segmented into a set of word regions. Then, word shape features are extracted and stored. Given a user's query with text, features are extracted after its corresponding word image is generated. This reference model is compared with the stored features to find out similar words. The proposed system is implemented with IBM-PC in a web environment and its experiments are performed with English document images. Experimental results show the effectiveness of the proposed methods.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.1 s.307
• /
• pp.53-66
• /
• 2006

This paper proposes the combined image retrieval system that gives the same relevance as exhaustive search method while its performance can be considerably improved. This system is combined with two different retrieval methods and each gives the same results that full exhaustive search method does. Both of them are two-stage method. One uses condensation of feature vectors, and the other uses binary-tree clustering. These two methods extract the candidate images that always include correct answers at the first stage, and then filter out the incorrect images at the second stage. Inasmuch as these methods use equal algorithm, they can get the same result as full exhaustive search. The first method condenses the dimension of feature vectors, and it uses these condensed feature vectors to compute similarity of query and images in database. It can be found that there is an optimal condensation ratio which minimizes the overall retrieval time. The optimal ratio is applied to first stage of this method. Binary-tree clustering method, searching with recursive 2-means clustering, classifies each cluster dynamically with the same radius. For preserving relevance, its range of query has to be compensated at first stage. After candidate clusters were selected, final results are retrieved by computing similarities again at second stage. The proposed method is combined with above two methods. Because they are not dependent on each other, combined retrieval system can make a remarkable progress in performance.

• Journal of Internet Computing and Services
• /
• v.19 no.6
• /
• pp.91-99
• /
• 2018

Sketch-based 3D object retrieval is a convenient way to search for various 3D data using human-drawn sketches as query. In this paper, we propose a new method of using Sketch CNN, Wasserstein CNN and Wasserstein center loss for sketch-based 3D object search. Specifically, Wasserstein center loss is a method of learning the center of each object category and reducing the Wasserstein distance between center and features of the same category. To do this, the proposed 3D object retrieval is performed as follows. Firstly, Wasserstein CNN extracts 2D images taken from various directions of 3D object using CNN, and extracts features of 3D data by computing the Wasserstein barycenters of features of each image. Secondly, the features of the sketch are extracted using a separate Sketch CNN. Finally, we learn the features of the extracted 3D object and the features of the sketch using the proposed Wasserstein center loss. In order to demonstrate the superiority of the proposed method, we evaluated two sets of benchmark data sets, SHREC 13 and SHREC 14, and the proposed method shows better performance in all conventional metrics compared to the state of the art methods.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.34 no.3
• /
• pp.329-338
• /
• 2016

As the need of indoor spatial information has grown, many applications have been developed. Nevertheless, the major representations of indoor spatial information are on the 2D or 3D, recently, the service based on omni-directional image has increased. Current service based on omni-directional image is used just for viewer. To provide various applications which can serve the identifying the attribute of indoor space, query based services and so on, topological data which can define the spatial relationships between spaces is required. For developing diverse applications based on omni-directional image, this study proposes the method to generate IndoorGML data which is the international standard of indoor topological data model. The proposed method is consist of 3 step to generate IndoorGML data; 1) Analysis the core elements to adopt IndoorGML concept to image, 2) Propose the method to identify the element of ‘Space’ which is the core element of IndoorGML concept, 3) Define the connectivity of indoor spaces. The proposed method is implemented at the 6-floor of 21centurybuilding of the University of Seoul to generate IndoorGML data and the demo service is implemented based on the generated data. This study has the significance to propose a method to generate the indoor topological data for the indoor spatial information services based on the IndoorGML.

• Journal of the Korean Society for Library and Information Science
• /
• v.44 no.1
• /
• pp.101-115
• /
• 2010

The main difference between web searching and traditional searching is that the web provides and supports multimedia searching. This study aims to investigate the multimedia searching behavior of users of NAVER, a major Korean search portal. In conducting this study, the query logs and click logs of a unified search service were analyzed. The results of this study show that among the multimedia queries submitted by users, audio searches are the dominant media type, followed similarly by video and image searches. On the other hand, among the multimedia documents clicked on, video is the most popular collection type followed by image and audio collections. Entertainment is the most popular topic in both multimedia queries and clicks. The results of this study can be implemented for the portal's development of multimedia content and searching algorithms.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.38 no.5
• /
• pp.58-64
• /
• 2001

Recently many informations are transmitted ,md stored as video data, and they are on the rapid increase because of popularization of high performance computer and internet. In this paper, to retrieve video data, shots are found through analysis of video stream and the method of detection of key frame is studied. Finally users can retrieve the video efficiently. This Paper suggests a new feature that is robust to object movement in a shot and is not sensitive to change of color in boundary detection of shots, and proposes the characterizing value that reflects the characteristic of kind of video (movie, drama, news, music video etc,). The key frames are pulled out from many frames by using the local minima and maxima of differential of the value. After original frame(not de image) are reconstructed for key frame, indexing process is performed through computing parameters. Key frames that arc similar to user's query image arc retrieved through computing parameters. It is proved that the proposed methods are better than conventional method from experiments. The retrieval accuracy rate is so high in experiments.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.195-206
• /
• 2022

Crack detection is essential for inspection of existing structures and crack segmentation based on deep learning is a significant solution. However, datasets are usually one of the key issues. When building a new dataset for deep learning, laborious and time-consuming annotation of a large number of crack images is an obstacle. The aim of this study is to develop an approach that can automatically select a small portion of the most informative crack images from a large pool in order to annotate them, not to label all crack images. An active learning method with difficulty learning mechanism for crack segmentation tasks is proposed. Experiments are carried out on a crack image dataset of a steel box girder, which contains 500 images of 320×320 size for training, 100 for validation, and 190 for testing. In active learning experiments, the 500 images for training are acted as unlabeled image. The acquisition function in our method is compared with traditional acquisition functions, i.e., Query-By-Committee (QBC), Entropy, and Core-set. Further, comparisons are made on four common segmentation networks: U-Net, DeepLabV3, Feature Pyramid Network (FPN), and PSPNet. The results show that when training occurs with 200 (40%) of the most informative crack images that are selected by our method, the four segmentation networks can achieve 92%-95% of the obtained performance when training takes place with 500 (100%) crack images. The acquisition function in our method shows more accurate measurements of informativeness for unlabeled crack images compared to the four traditional acquisition functions at most active learning stages. Our method can select the most informative images for annotation from many unlabeled crack images automatically and accurately. Additionally, the dataset built after selecting 40% of all crack images can support crack segmentation networks that perform more than 92% when all the images are used.